We have two main classes [class eNB, class UE] that store the information related to both the cellular tower and user equipment. The next step involved is creating a reasonable probability transition matrix within a grid of a NxN dimensions. In our implementation, N = 500 miles. The probabilities were chosen based on their distances to each other, giving the closest stations to one node a higher probability. The dimensions of the matrix is simply MxM, where M = number of stations within the grid. The initial distribution of each station is instantiated with respect to each user that is in the grid. The values are also chosen with respect to the location of the user to the other stations, the speed of the user, and the Signal-to-Noise ratio of the channel also with respect to the user and a certain eNB station. After the transition matrix and the initial distribution have been created, the simulation is run using the Markov transition property for each time step [one hour] into the future, generating a newer distribution for each station as time passes. The predicted station is simply the maximum probability value of the distribution.